-
Notifications
You must be signed in to change notification settings - Fork 4
/
Copy pathinference.py
181 lines (148 loc) · 6.67 KB
/
inference.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
# coding:utf-8
"""
USEAGE:
CUDA_VISIBLE_DEVICES=7 python inference.py -i=dataset/weibo/stc_weibo_train_post -p=exp/weibo/pretrain/2018-01-29-11-43-34/ -e=1
"""
import os
import pickle
import argparse
import numpy as np
from data import batcher, load_vocab, sentence2id, id2sentence, padding_inputs
from toy import reload_model
import torch
import torch.nn as nn
from torch.autograd import Variable
from encoder import EncoderRNN
from generator import Generator
from utils import SYM_PAD, SYM_EOS, SYM_GO, SYM_UNK
USE_CUDA = True
def predict(test_post_file, max_len, vocab, rev_vocab, word_embeddings, encoder, generator, output_file=None):
# data generator
test_data_generator = batcher(1, test_post_file, response_file=None)
if output_file:
fo = open(output_file, 'wb')
while True:
try:
post_sentence = test_data_generator.next()
except StopIteration:
logger.info('---------------------finish-------------------------')
break
post_ids = [sentence2id(sent, vocab) for sent in post_sentence]
posts_var, posts_length = padding_inputs(post_ids, None)
if USE_CUDA:
posts_var = posts_var.cuda()
embedded_post = word_embeddings(posts_var)
_, dec_init_state = encoder(embedded_post, input_lengths=posts_length.numpy())
log_softmax_outputs = generator.inference(dec_init_state, word_embeddings) # [B, T, vocab_size]
hyps, _ = beam_search(dec_init_state, max_len, word_embeddings, generator, beam=5, penalty=1.0, nbest=1)
results = []
for h in hyps:
results.append(id2sentence(h[0], rev_vocab))
print('*******************************************************')
print "post:" + ''.join(post_sentence[0])
print "response:\n" + '\n'.join([''.join(r) for r in results])
print
# if algo == 'greedy':
# # or 'beam'
# _, results = torch.max(log_softmax_outputs, dim=2) # [B, T]
# elif algo == 'beam':
# hyps, _ = beam_search(dec_init_state, max_len, word_embeddings, generator)
#
# response_sentence = id2sentence(results.cpu().data.numpy().reshape(-1).tolist(), rev_vocab)
# if not output_file:
# print('*******************************************************')
# print "post:" + ''.join(post_sentence[0])
# print "response:" + ''.join(response_sentence)
# else:
# fo.write("post: %s\nresponse: %s\n\n" % (
# ''.join(post_sentence[0]),
# '',join(response_sentence)))
# if output_file:
# fo.close()
def beam_search(dec_init_state, max_len, word_embeddings, generator, beam=5, penalty=1.0, nbest=1, use_cuda=True):
"""
the code is referred by:
https://github.com/dialogtekgeek/DSTC6-End-to-End-Conversation-Modeling/blob/master/ChatbotBaseline/tools/seq2seq_model.py
dec_init_state comes from encoder
"""
# input [B, T, emb_dim]
go_i = Variable(torch.LongTensor([[SYM_GO]]), requires_grad=False)
eos_i = Variable(torch.LongTensor([[SYM_EOS]]), requires_grad=False)
if use_cuda:
go_i = go_i.cuda()
eos_i = eos_i.cuda()
ds = generator.update((None, dec_init_state), word_embeddings(go_i))
hyplist = [([], 0., ds)]
best_state = None
comp_hyplist = []
for l in range(max_len):
new_hyplist = []
argmin = 0
for out, lp, st in hyplist:
logp = generator.predict(st, word_embeddings)
#[vocab_size,]
lp_vec = logp.cpu().data.numpy()[0] + lp
if l > 0:
new_lp = lp_vec[SYM_EOS] + penalty*(len(out)+1)
new_st = generator.update(st, word_embeddings(eos_i))
comp_hyplist.append((out, new_lp))
if best_state is None or best_state[0] < new_lp:
best_state = (new_lp, new_st)
for o in np.argsort(lp_vec)[::-1]:
if o == SYM_UNK or o == SYM_EOS:
continue
new_lp = lp_vec[o]
if len(new_hyplist) == beam:
if new_hyplist[argmin][1] < new_lp:
new_st = generator.update(st,
word_embeddings(Variable(torch.LongTensor([[o]]), requires_grad=False).cuda()))
new_hyplist[argmin] = (out+[o], new_lp, new_st)
argmin = min(enumerate(new_hyplist), key=lambda h:h[1][1])[0]
else:
break
else:
new_st = generator.update(st,
word_embeddings(Variable(torch.LongTensor([[o]]), requires_grad=False).cuda()))
new_hyplist.append((out+[o], new_lp, new_st))
if len(new_hyplist) == beam:
argmin = min(enumerate(new_hyplist), key=lambda h:h[1][1])[0]
hyplist = new_hyplist
if len(comp_hyplist):
maxhyps = sorted(comp_hyplist, key=lambda h: -h[1])[:nbest]
return maxhyps, best_state[1]
else:
return [([], 0)], None
def main():
argparser = argparse.ArgumentParser()
argparser.add_argument('--test_query_file', '-i', type=str, required=True)
argparser.add_argument('--load_path', '-p', type=str, required=True)
# TODO: load epoch -> load best model
argparser.add_argument('--load_epoch', '-e', type=int, required=True)
argparser.add_argument('--output_file', '-o', type=str)
argparser.add_argument('--dec_algorithm', '-algo', type=str, default='greedy')
new_args = argparser.parse_args()
arg_file = os.path.join(new_args.load_path, 'args.pkl')
if not os.path.exists(arg_file):
raise RuntimeError('No default arguments file to load')
f = open(arg_file, 'rb')
args = pickle.load(f)
f.close()
if args.use_cuda:
USE_CUDA = True
vocab, rev_vocab = load_vocab(args.vocab_file, max_vocab=args.max_vocab_size)
vocab_size = len(vocab)
word_embeddings = nn.Embedding(vocab_size, args.emb_dim, padding_idx=SYM_PAD)
E = EncoderRNN(vocab_size, args.emb_dim, args.hidden_dim, args.n_layers, bidirectional=True, variable_lengths=True)
G = Generator(vocab_size, args.response_max_len, args.emb_dim, 2*args.hidden_dim, args.n_layers)
if USE_CUDA:
word_embeddings.cuda()
E.cuda()
G.cuda()
reload_model(new_args.load_path, new_args.load_epoch, word_embeddings, E, G)
predict(new_args.test_query_file,
args.response_max_len,
vocab, rev_vocab,
word_embeddings, E, G,
new_args.output_file)
if __name__ == '__main__':
main()